Environmental Factor, April 2008, National Institute of Environmental Health Sciences
New Directions in Autism Research Funded by NIEHS
By Eddy Ball
Support from NIEHS is enabling researchers at NIEHS and across the country make important discoveries that may help unravel the mysteries of the complex syndromes grouped as Autism Spectrum Disorders (ASD). New studies have been published in the past few months by NIEHS-funded investigators involved with the Childhood Autism Risks from Genetics and Environment (CHARGE) project and their colleagues. To promote further collaborative efforts in the field, NIEHS is also participating in the Interagency Autism Coordination Committee on Autism Research(http://www.nimh.nih.gov/research-funding/scientific-meetings/recurring-meetings/iacc/index.shtml) .
One of these new studies found an increase in plasma levels of the hormone/cytokine leptin in children with early onset and regressive autism that may be the first biological marker that can distinguish the two forms of the disorder. A second study identified immunological factors in mothers that could be linked to autism in the very earliest stages of life. A third study identified for the first time specific patterns of gene expression in autistic children that differentiate them from normally developing children in the general population. The researchers were also able to distinguish between autistic children with early onset or regressive autism using gene expression patterns.
A new study is about to be launched with support from NIEHS to identify very early environmental exposures that might contribute to the development of autism. The Early Autism Risk Longitudinal Investigation (EARLI) Network aims to follow 1,200 mothers of children with autism at the start of a new pregnancy and document the development of their newborn siblings through age three.
This study will provide a unique opportunity for studying possible environmental risk factors for autism and biomarkers during different developmental windows, as well as an opportunity to investigate the interplay of genetic susceptibility and environmental exposure. A number of such exposures, ranging from suspected neurotoxicants to medications taken during pregnancy, could potentially be investigated with data and samples collecting through EARLI.
There has been significant public interest in the potential role of the vaccine preservative thimerosal in the recent increase in the prevalence of autism spectrum disorders. In an effort to replicate findings reported in a 2004 publication on the strain-dependent neurodevelopmental effects in mice from thimerosal exposure, a group of NIEHS-funded researchers repeated the experiment. They applied rigorous methodology to enable collection of additional toxicological endpoints and behavioral assessments. The investigators' systematic and objective approach added important and reliable scientific evidence to the ongoing discourse about the effects of the preservative.
Taking CHARGE of Autism Research
Initiated in 2002, the CHARGE(http://beincharge.ucdavis.edu/) Study represents the first large-scale, population-based and case-control epidemiologic study of autism. The study is based at the University of California (UC) Davis, housed at the university's Medical Investigation of Neurological Disorders (MIND)(http://www.ucdmc.ucdavis.edu/mindinstitute/) Institute and linked to the laboratories at its Center for Children's Environmental Health. CHARGE studies have addressed a wide spectrum of environmental exposures and genetic and epigenetic susceptibility factors and their interplay. In 2003, the study set out to recruit 1,000 to 2,000 children with differing patterns of development. Investigators sought to enroll a population of children with autism, children with developmental delay who do not have autism, and children from the general population. In 2006, the study included 1,240 subjects.
CHARGE studies have explored the phenotypic heterogeneity of Autistic Spectrum Disorders (ASD) in an effort to understand the extent to which several conditions with distinct etiologies and pathogenic mechanisms may be involved. Researchers have investigated similarities and differences in developmental delays in affected children and how those delays skew the incidence ratio of four boys to one girl often observed in ASD. Exploration is underway of co-morbidity patterns in ASD, such as gastrointestinal problems and sleep disturbances.
In an attempt to delineate the mechanisms of pathogenesis, CHARGE investigators have utilized extensive interviews with parents, laboratory analysis of blood, urine and hair specimens, and prenatal labor and delivery, neonatal and pediatric records. The research also takes advantage of animal models of autism and in vitro investigations of immune and neurogenic cells.
Over the years, CHARGE has worked to build an infrastructure that can support multiple investigations of autism and related neurodevelopmental disorders. Because of the large number of cases of autism in the CHARGE study, it may be comparable with what can be expected in unselected birth cohorts of 100,000. Collaborations with other population-based efforts are currently under way, such as the national Centers for Autism and Developmental Disabilities Research and Epidemiology (CADDRE) study, which offer potential opportunities for replication and data pooling.
Citation: Hertz-Picciotto I, Croen LA, Hansen R, Jones CR, van de Water J, Pessah IN.(http://www.ncbi.nlm.nih.gov/pubmed/16835068?ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum) 2004. The CHARGE study: an epidemiologic investigation of genetic and environmental factors contributing to autism. Environ Health Perspect 114(7):1119-1125.
Elevated Leptin and Inflammation in Autism
Most of the hormone leptin is produced in fat cells. This has led to the association of the neuroendocrine mediator with regulation of energy intake and energy expenditure, including the regulation (decrease) of appetite and (increase) of metabolism, and hence to overweight and obesity conditions.
However, as the study by UC Davis immunologist Judy Van de Water, Ph.D., and her colleagues explains, there is evidence from earlier studies to demonstrate the production of leptin also by inflammatory cells. Researchers have shown in animal studies that leptin deficiencies can shift the immune response and result in defective cell-mediated and humoral immunity.
Working on the hypothesis that "an altered immune response may impact other biological systems including the neuroendocrine and nervous systems," the investigators compared plasma leptin levels, measured as nanogram/milliliter (ng/ml), in a population of children with careful documentation of autism and age-matched controls. Controls included children with and without developmental disabilities. Because of the association between body mass index (BMI) and leptin, the research team controlled for BMI-for-age, a modified calculation that is a more relevant measure for a pediatric population.
While the researchers found that leptin levels were significantly higher in children with autism than in typically developing controls, they also discovered a dramatic difference among the autistic children themselves. When the subjects with autism were further categorized by symptom cluster, children with early onset autism showed significantly higher median and interquartile range levels of leptin than did children with clinical regression: 2.62 (1.12 - 4.38) versus 1.38 (.061 - 2.69) ng/ml. The investigators also found that children with the regressive form of autism were not significantly different when compared with controls, both typically developing and developmentally delayed.
This study is the first report of a potential biochemical marker in autism that may be able to differentiate between disease phenotypes. The investigators acknowledged that it is still undetermined whether leptin levels are a cause or secondary phenomenon of the condition. However, they concluded, "these findings provide a framework for further longitudinal studies to investigate changes in leptin levels over the lifetime of the disorder."
Citation: Ashwood P, Kwong C, Hansen R, Hertz-Picciotto I, Croen L, Krakowiak P, Walker W, Pessah IN, Van de Water J.() 2008. Brief Report: Plasma Leptin Levels are Elevated in Autism: Association with Early Onset Phenotype? J Autism Dev Disord 38(1):169-175.
IgG Antibody Transfer and Early Neurodevelopment
The search for a biochemical marker of autism moved forward with another recent investigation of maternal plasma IgG antibodies against human fetal and adult brain proteins, also led by Van de Water. According to the investigators, the antibodies, which can easily cross the placental barrier, may be able to react to fetal 'self'-proteins to give the fetus a subset of the maternal adaptive humoral immune system proteins. These proteins could possibly produce an autoimmune reaction in the infant which could then impact fetal neurodevelopment.
Researchers studied 61 mothers of children with autistic disorder (AU) and 102 matched controls with typical developing (TD) children or children with non-ASD developmental delays (DD). Separating proteins by weight, they looked for patterns of reactivity in the range of 20 to 220 kilo-daltons (kDa) - a measure of molecular weight used in chemiluminescent visualization of marker bands.
The investigators found patterns of reactivity against fetal, but not adult, brain clustered at two points, 37kDa and 73kDa, that differentiated between mothers of AD children and mothers of TD and DD children. Autoreactivity to a protein at approximately 37kDa was found in 26 percent of mothers of AD children compared to 2.5 percent of mothers of DD children and 8.1 percent of mothers with TD children, yielding a 5.69-fold odds ratio associated with this band.
When reactivity against proteins at both molecular weights was studied, 11 percent of mothers of AU children showed the results at 37kDa and 73kDa, while no mothers of TD or DD children showed that pattern of reactivity. Significantly, 86 percent of the mothers of AU children with reactivity to the pair of bands had children with the regressive form of the disease.
The team found support for their hypothesis that there is a potential role for maternal IgG antibody transplacental transfer during pregnancy and subsequent binding to fetal brain cells as a cause of autism in some children. They also discovered a possible biomarker that may be able to predict the development of a specific phenotype of the disease.
The results led Van de Water to comment, "We are ...optimistic that in the future a prenatal test and therapeutic intervention preventing IgG exposure during pregnancy could protect some children from ever getting autism."
Citation: Braunschweig D, Ashwood P, Krakowiak P, Hertz-Picciotto I, Hansen R, Croen LA, Pessah IN, Van de Water J.() Autism: Maternally derived antibodies specific for fetal brain proteins. Neurotoxicology epub 2007
Gene Expression Patterns in Autistic Children
A precedent-setting CHARGE investigation of gene expression patterns in autistic children could give researchers new insights into the biological foundations of the disease and may be an important step in identifying new targets for therapies.
Led by UC Davis neurologist Frank Sharp, M.D., the investigation was the first to use genomic profiling of whole blood successfully to pinpoint differential gene expression patterns that distinguished autistic (AU) children, diagnosed by both the Autism Diagnostic Observation Schedule and the Autism Diagnostic Interview, from general population (GP) children. The team also was the first to identify gene expression patterns that differed between children with the major phenotypes or subsets of AU, the early onset without regression (A-E) and the regressive forms (A-R) of the disease.
The subject population included 61 children enrolled in the ongoing CHARGE study: 35 diagnosed with AU; 14 diagnosed with Autism Spectrum Disorder (ASD), a term used for children who do not meet the full behavioral criteria for AU; and 12 typically developing GP children, age and gender matched, with no evidence of ASD. RNA was isolated from blood drawn in the afternoon and analyzed for gene expression patterns.
Results indicated that the autistic groups (AU, A-R, and A-E) shared 11 differentially expressed genes in natural killer (NK) cells, many of which belong to the natural-killer cytotoxicity pathway identified in the Kyoto Encyclopedia of Genes and Genomes. The authors observed that "the importance of these findings is that NK cells are a primary, innate defense against viral, bacterial, and parasitic infections or malignant transformation."
A secondary finding, but one of importance in the diagnosis of forms of AU, involves differentially expressed gene patterns that differ between A-E and GP children, a set of 140 genes, and between A-R and GP children, a set of 20 genes. Gene expression data supporting different forms of AU, the authors concluded, "may be able to help define the etiology, genetics and clinical phenotype, as well as the outcome, in autism."
Gregg JP, Lit L, Baron CA, Hertz-Picciotto I, Walker W, Davis RA, Croen LA, Ozonoff S, Hansen R, Pessah IN, Sharp FR.() 2008. Gene expression changes in children with autism. Genomics 91(1):22-29.
Mice Models of Thimerosal Toxicity
In the long and often frustrating search for what environmental factors might trigger autism (AU) and autistic spectrum disorders (ASD), concerned parents and some researchers have speculated that vaccine-level exposure to ethylmercury in thimerosal-preserved childhood vaccines could play a role. Advocates of the thimerosal connection hypothesized that deficits in immune function among some children may make their developing nervous systems vulnerable to pervasive developmental toxicity, leading to AU and ASD.
While relevant data in humans are lacking, a 2004 study() of the autoimmune-susceptible SJL/J mice strain conducted by Mady Hornig, M.D., and associates at Columbia University's Mailman School of Public Health, reported findings consistent with the notion that immune dysfunction can exacerbate the effects of thimerosal. Their study found gross morphological changes in the hippocampus as a result of early post-natal injections of thimerosal.
A research team that included NIEHS Neurotoxicology Group Head G. Jean Harry, Ph.D., attempted to replicate Hornig's findings, given the significance of this approach of an augmented response to toxicants as a function of underlying immune status. The team consisted of researchers experienced in the design and conduct of developmental neurotoxicology studies and with expertise in morphological and statistical analysis. The animal studies were conducted in laboratories at the University of California Davis and used a similar injection schedule to that reported by Hornig. SJL/J pups from 51 different litters were injected with five different solutions according to a regimen that modeled childhood vaccination schedules.
A within-litter dosing design was utilized to control for the large impact that maternal factors, both genetic and behavioral, have on the neurobehavioral functioning of the offspring. Tissue mercury levels were used to confirm that there was no cross contamination within the litters from the pups injected with thimerosal.
The first group received a dose of thimerosal (1X thimerosal) comparable to the maximum dosage a child could have been exposed to with thimerosal-preserved vaccines against hepatitis B, diphtheria tetanus pertussis (DPT) and hemophilus influenza B (HiB). A control group received a phosophate-buffered saline vehicle (VEH). The other groups received, respectively, vaccine alone (VAC), 1X + VAC, and 10X + VAC to evaluate the effects of different levels of thimerosal and the different components of vaccines.
After the mice had completed the injection schedule, the research team evaluated survival, body weight growth indices of early development and hippocampal morphology. They put the mice through a battery of behavioral tests related to core neurological deficits, including social interaction, sensory gating and anxiety.
In one cohort of mice, total mercury levels were measured in blood, brain and kidney at two time points selected based upon previous work by the UC Davis lab and Harry lab at NIEHS. Morphological examination of the hippocampus was conducted using an advanced computerized methodology of unbiased stereology performed by an expert in the field. In all cases, data collection and analysis were conducted on coded animals and samples to maintain a totally unbiased assessment with investigators blind to individual animal treatment groups. The code was maintained during the statistical analysis as well.
In contrast to the Hornig study, the mice in this study maintained overall good health and showed no adverse behaviors. The researchers found no evidence of structural abnormalities in the hippocampus or of neurobehavioral changes in activity levels or social behavior in the offspring as a result of the injections of thimerosal, by itself or in combination with VAC. As expected, pups injected with the higher 10X and 10X + VAC showed higher blood levels of mercury; however, the other results were similar to the lower exposure groups.
Berman RF, Pessah IN, Mouton PR, Mav D, Harry J.(http://www.ncbi.nlm.nih.gov/pubmed/17977901?ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum) 2008. Low-Level Neonatal Thimerosal Exposure: Further Evaluation of Altered Neurotoxic Potential in SJL Mice. Toxicol Sci 101(2):294-309.
Future Directions: Networking for a Large-Scale Epidemiological Study of Gene-Environment Interactions
Thepromises to pull together several lines of research from earlier CHARGE investigations and explore new directions in a ten-year study of mothers of autistic children and their newborn siblings that combines epidemiology and basic laboratory research. The network includes an administrative center at the Drexel University School of Public Health, a data coordinating center at the University of California Davis, a central lab and repository at Johns Hopkins School of Public Health, and field sites in Philadelphia, Baltimore, the San Francisco Bay area and Davis, Calif.
The network plans to implement a core epidemiologic data collection protocol focusing on prospective documentation of exogenous exposures, continuous ASD behavioral domains during pregnancy and early life, collection and banking of biological samples and follow-up of the newborn siblings through 36 months of age. Researchers will gather data on ASD diagnoses, continuous ASD behavioral domains and other behaviors that may be associated with ASD.
Principal Investigator Craig J. Newschaffer, Ph.D., of Drexel University, listed four "exemplary specific aims" to be addressed in the study:
- Determine whether markers of maternal autoimmune status, measured during pregnancy, at delivery and at six months post-partum, are associated with autism risk
- Determine whether in utero exposure to persistent organic pollutants is associated with autism risk
- Explore maternal and child epigenetic marks as predictors of ASD risk
- Assess whether polymorphisms of ADRB2, which may affect brain development, and in utero ADRB2 agonist exposure are associated with autism risk and explore potential interaction of genotype and exposure
- Data collected in this large study are intended to serve as a resource for ongoing epidemiologic investigation of potential risk factors and risk biomarkers for ASD well into the future.